Explainable Artificial Intelligence (XAI)

Derniers papiers

• Step-Level Sparse Autoencoder for Reasoning Process Interpretation

  Xuan Yang, Jiayu Liu, Yuhang Lai, Hao Xu, Zhenya Huang, Ning Miao · 4 mars 2026

  Large Language Models (LLMs) have achieved strong complex reasoning capabilities through Chain-of-Thought (CoT) reasoning. However, their reasoning patterns remain too complicated to analyze. While Sparse Autoencoders (SAEs) have emerged as a powerful tool for interpretability, existing approaches p…

• Beyond Factual Correctness: Mitigating Preference-Inconsistent Explanations in Explainable Recommendation

  Chengkai Wang, Baisong Liu · 4 mars 2026

  LLM-based explainable recommenders can produce fluent explanations that are factually correct, yet still justify items using attributes that conflict with a user's historical preferences. Such preference-inconsistent explanations yield logically valid but unconvincing reasoning and are largely misse…

• Contextual Invertible World Models: A Neuro-Symbolic Agentic Framework for Colorectal Cancer Drug Response

  Christopher Baker, Karen Rafferty, Hui Wang · 4 mars 2026

  Precision oncology is currently limited by the small-N, large-P paradox, where high-dimensional genomic data is abundant, but high-quality drug response samples are often sparse. While deep learning models achieve high predictive accuracy, they remain black boxes that fail to provide the causal mech…

• Exact Functional ANOVA Decomposition for Categorical Inputs Models

  Baptiste Ferrere (IMT, SINCLAIR AI Lab), Nicolas Bousquet (SINCLAIR AI Lab), Fabrice Gamboa (IMT, ANITI), Jean-Michel Loubes (IMT, REGALIA, ANITI), Joseph Mur\'e · 4 mars 2026

  Functional ANOVA offers a principled framework for interpretability by decomposing a model's prediction into main effects and higher-order interactions. For independent features, this decomposition is well-defined, strongly linked with SHAP values, and serves as a cornerstone of additive explainabil…

• CoBELa: Steering Transparent Generation via Concept Bottlenecks on Energy Landscapes

  Sangwon Kim, Kyoungoh Lee, Jeyoun Dong, Kwang-Ju Kim · 4 mars 2026

  Generative concept bottleneck models aim to enable interpretable generation by routing synthesis through explicit, user-facing concepts. In practice, prior approaches often rely on non-explicit bottleneck representations (e.g., vision cues or opaque concept embeddings) or black-box decoders to prese…

• Neuro-Symbolic Artificial Intelligence: A Task-Directed Survey in the Black-Box Models Era

  Giovanni Pio Delvecchio, Lorenzo Molfetta, Gianluca Moro · 4 mars 2026

  The integration of symbolic computing with neural networks has intrigued researchers since the first theorizations of Artificial intelligence (AI). The ability of Neuro-Symbolic (NeSy) methods to infer or exploit behavioral schema has been widely considered as one of the possible proxies for human-l…

• Why Does RLAIF Work At All?

  Robin Young · 4 mars 2026

  Reinforcement Learning from AI Feedback (RLAIF) enables language models to improve by training on their own preference judgments, yet no theoretical account explains why this self-improvement seemingly works for value learning. We propose the latent value hypothesis, that pretraining on internet-sca…

• Decoding Answers Before Chain-of-Thought: Evidence from Pre-CoT Probes and Activation Steering

  Kyle Cox, Darius Kianersi, Adri\`a Garriga-Alonso · 3 mars 2026

  As chain-of-thought (CoT) has become central to scaling reasoning capabilities in large language models (LLMs), it has also emerged as a promising tool for interpretability, suggesting the opportunity to understand model decisions through verbalized reasoning. However, the utility of CoT toward inte…

• Extended Empirical Validation of the Explainability Solution Space

  Antoni Mestre, Manoli Albert, Miriam Gil, Vicente Pelechano · 3 mars 2026

  This technical report provides an extended validation of the Explainability Solution Space (ESS) through cross-domain evaluation. While initial validation focused on employee attrition prediction, this study introduces a heterogeneous intelligent urban resource allocation system to demonstrate the g…

• CARE: Confounder-Aware Aggregation for Reliable LLM Evaluation

  Jitian Zhao, Changho Shin, Tzu-Heng Huang, Satya Sai Srinath Namburi GNVV, Frederic Sala · 3 mars 2026

  LLM-as-a-judge ensembles are the standard paradigm for scalable evaluation, but their aggregation mechanisms suffer from a fundamental flaw: they implicitly assume that judges provide independent estimates of true quality. However, in practice, LLM judges exhibit correlated errors caused by shared l…

• LOCUS: A Distribution-Free Loss-Quantile Score for Risk-Aware Predictions

  Matheus Barreto, M\'ario de Castro, Thiago R. Ramos, Denis Valle, Rafael Izbicki · 3 mars 2026

  Modern machine learning models can be accurate on average yet still make mistakes that dominate deployment cost. We introduce Locus, a distribution-free wrapper that produces a per-input loss-scale reliability score for a fixed prediction function. Rather than quantifying uncertainty about the label…

• Generalizing Logic-based Explanations for Machine Learning Classifiers via Optimization

  Francisco Mateus Rocha Filho, Ajalmar R\^ego da Rocha Neto, Thiago Alves Rocha · 3 mars 2026

  Machine learning models support decision-making, yet the reasons behind their predictions are opaque. Clear and reliable explanations help users make informed decisions and avoid blindly trusting model outputs. However, many existing explanation methods fail to guarantee correctness. Logic-based app…

• EigenBench: A Comparative Behavioral Measure of Value Alignment

  Jonathn Chang, Leonhard Piff, Suvadip Sana, Jasmine X. Li, Lionel Levine · 3 mars 2026

  Aligning AI with human values is a pressing unsolved problem. To address the lack of quantitative metrics for value alignment, we propose EigenBench: a black-box method for comparatively benchmarking language models' values. Given an ensemble of models, a constitution describing a value system, and …

• Hidden Breakthroughs in Language Model Training

  Sara Kangaslahti, Elan Rosenfeld, Naomi Saphra · 3 mars 2026

  Loss curves are smooth during most of model training, so visible discontinuities stand out as possible conceptual breakthroughs. Studying these breakthroughs enables a deeper understanding of learning dynamics, but only when they are properly identified. This paper argues that similar breakthroughs …

• Designing Explainable AI for Healthcare Reviews: Guidance on Adoption and Trust

  Eman Alamoudi, Ellis Solaiman · 3 mars 2026

  Patients increasingly rely on online reviews when choosing healthcare providers, yet the sheer volume of these reviews can hinder effective decision-making. This paper summarises a mixed-methods study aimed at evaluating a proposed explainable AI system that analyses patient reviews and provides tra…

• Bound Propagation meets Constraint Simplification: Improving Logic-based XAI for Neural Networks

  Ronaldo Gomes, Jairo Ribeiro, Luiz Queiroz, Thiago Alves Rocha · 3 mars 2026

  Logic-based methods for explaining neural network decisions offer formal guarantees of correctness and non-redundancy, but they often suffer from high computational costs, especially for large networks. In this work, we improve the efficiency of such methods by combining bound propagation with const…

• Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning

  Hamed Damirchi, Ignacio Meza De la Jara, Ehsan Abbasnejad, Afshar Shamsi, Zhen Zhang, Javen Shi · 3 mars 2026

  Existing explainability methods for Large Language Models (LLMs) typically treat hidden states as static points in activation space, assuming that correct and incorrect inferences can be separated using representations from an individual layer. However, these activations are saturated with polyseman…

• What Helps -- and What Hurts: Bidirectional Explanations for Vision Transformers

  Qin Su, Tie Luo · 3 mars 2026

  Vision Transformers (ViTs) achieve strong performance in visual recognition, yet their decision-making remains difficult to interpret. We propose BiCAM, a bidirectional class activation mapping method that captures both supportive (positive) and suppressive (negative) contributions to model predicti…

• Jump Like A Squirrel: Optimized Execution Step Order for Anytime Random Forest Inference

  Daniel Biebert, Christian Hakert, Kay Heider, Daniel Kuhse, Sebastian Buschj\"ager, Jian-Jia Chen · 3 mars 2026

  Due to their efficiency and small size, decision trees and random forests are popular machine learning models used for classification on resource-constrained systems. In such systems, the available execution time for inference in a random forest might not be sufficient for a complete model execution…

• Explanation-Guided Adversarial Training for Robust and Interpretable Models

  Chao Chen, Yanhui Chen, Shanshan Lin, Dongsheng Hong, Shu Wu, Xiangwen Liao, Chuanyi Liu · 3 mars 2026

  Deep neural networks (DNNs) have achieved remarkable performance in many tasks, yet they often behave as opaque black boxes. Explanation-guided learning (EGL) methods steer DNNs using human-provided explanations or supervision on model attributions. These approaches improve interpretability but typi…

• I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data?

  Yuhang Liu, Dong Gong, Yichao Cai, Erdun Gao, Zhen Zhang, Biwei Huang, Mingming Gong, Anton van den Hengel, Javen Qinfeng Shi · 3 mars 2026

  Recent empirical evidence shows that LLM representations encode human-interpretable concepts. Nevertheless, the mechanisms by which these representations emerge remain largely unexplored. To shed further light on this, we introduce a novel generative model that generates tokens on the basis of such …

• On The Fragility of Benchmark Contamination Detection in Reasoning Models

  Han Wang, Haoyu Li, Brian Ko, Huan Zhang · 3 mars 2026

  Leaderboards for LRMs have turned evaluation into a competition, incentivizing developers to optimize directly on benchmark suites. A shortcut to achieving higher rankings is to incorporate evaluation benchmarks into the training data, thereby yielding inflated performance, known as benchmark contam…

• ASTRA-bench: Evaluating Tool-Use Agent Reasoning and Action Planning with Personal User Context

  Zidi Xiu, David Q. Sun, Kevin Cheng, Maitrik Patel, Josh Date, Yizhe Zhang, Jiarui Lu, Omar Attia, Raviteja Vemulapalli, Oncel Tuzel, Meng Cao, Samy Bengio · 3 mars 2026

  Next-generation AI must manage vast personal data, diverse tools, and multi-step reasoning, yet most benchmarks remain context-free and single-turn. We present ASTRA-bench (Assistant Skills in Tool-use, Reasoning \& Action-planning), a benchmark that uniquely unifies time-evolving personal context w…

• Knowledge without Wisdom: Measuring Misalignment between LLMs and Intended Impact

  Michael Hardy, Yunsung Kim · 3 mars 2026

  LLMs increasingly excel on AI benchmarks, but doing so does not guarantee validity for downstream tasks. This study evaluates the performance of leading foundation models (FMs, i.e., generative pre-trained base LLMs) with out-of-distribution (OOD) tasks of the teaching and learning of schoolchildren…

• A Case Study on Concept Induction for Neuron-Level Interpretability in CNN

  Moumita Sen Sarma, Samatha Ereshi Akkamahadevi, Pascal Hitzler · 3 mars 2026

  Deep Neural Networks (DNNs) have advanced applications in domains such as healthcare, autonomous systems, and scene understanding, yet the internal semantics of their hidden neurons remain poorly understood. Prior work introduced a Concept Induction-based framework for hidden neuron analysis and dem…